How the slippage calculation works
An order book is a list of resting prices and the size displayed at each price. For a buy, the calculator sorts asks from the lowest price to the highest. For a sell, it sorts bids from the highest price to the lowest. It consumes the best level first, subtracts that size from the requested shares and continues until the request is filled or the entered ladder ends.
Volume-weighted average price, or VWAP, is total modeled cash divided by filled shares. Displayed slippage is the difference between that VWAP and the best entered price. A buy has positive slippage when its VWAP is above the best ask; a sell has positive slippage when its VWAP is below the best bid. The calculator reports the difference per share, as a cash total and in relative basis points.
Worked example: buying 1,000 shares
Suppose the entered asks show 800 shares at 0.52 and 1,500 at 0.53. A request for 1,000 shares consumes all 800 at 0.52 and another 200 at 0.53. Gross modeled cost is $522, VWAP is 0.522 and displayed slippage is 0.002 per filled share, or $2 across the request, before fees.
A top-of-book shortcut would price the entire request at 0.52 and understate cost. The error becomes larger as requested size grows relative to displayed depth. That is why a backtest based only on midpoint, last trade or best quote can look profitable while the executable ladder produces a weaker result.
How to interpret a Polymarket order book
Polymarket documents a central limit order book with bids and asks for outcome tokens. Its price guide distinguishes bid, ask, midpoint and last trade: a buyer crosses the ask and a seller crosses the bid. Its trading documentation also demonstrates walking price levels to calculate a market price for a requested amount. Paste asks into this tool for a buy and bids for a sell.
Use a book captured at the intended decision time. A current ladder cannot establish what was displayed yesterday, and a historical last trade is not a substitute for both sides of the historical book. Preserve token and market identifiers when moving observations into a backtest.
How to interpret a Kalshi order book
Kalshi describes its order book as resting offers with a price and number of contracts. A market order can consume several resting levels, so entered order size determines the average execution price. For a direct buy ladder, enter the prices you would pay from best to worse; for a sell, enter the bids that would receive the order.
Venue-native yes/no representations sometimes require a complement transform before comparison. Do that normalization upstream and confirm side semantics with the venue response. A numerically valid ladder on the wrong side of the contract will produce a precise but meaningless estimate.
Using slippage in a prediction-market backtest
Treat the calculator as a transparent single-observation audit. A robust backtest repeats the same ladder walk at every decision time and never loads a book observed after the signal. It also defines what happens when displayed size is insufficient: reject the trade, partially fill it, or carry a remaining order under an explicitly modeled rule.
- Select the latest eligible book observed at or before the decision timestamp.
- Reject stale books using a documented freshness threshold.
- Walk the correct side for the exact requested shares.
- Apply fees only under the venue, contract and order assumptions being studied.
- Store best price, VWAP, filled fraction, worst price and missing-depth reason.
- Compare depth-aware results with midpoint and fixed-slippage sensitivity cases.
What this calculator does not model
The result is a deterministic walk through the levels you enter. It does not model queue priority, cancellations between observation and arrival, network or matching-engine latency, hidden liquidity, price improvement, self-impact beyond the entered book, rejected orders or a fee schedule you did not enter. Displayed liquidity is evidence, not a guarantee of execution.
The examples are static teaching ladders, not current venue quotes. The tool does not connect to a wallet, submit an order or recommend a trade. For historical research, the quality of the estimate is bounded by the completeness, timing and identifier accuracy of the source order-book observation.